Baidu

According to Sina Technology, during Baidu's Q3 2024 earnings call, Robin Li, Baidu's Founder, Chairman and CEO, revealed that Baidu will release a new version of the Wenshin Big Model in early 2025 to consolidate its leadership in the underlying model. Robin Li said that thanks to the enhanced capabilities, widespread application and cost efficiency of the Wenshin Big Model, the average daily call volume of the Wenshin Big Model has increased significantly, reaching 1.5 billion times in November, and the average daily processing of Tokens text exceeded 1.7 trillion; compared with the 50 million daily calls first disclosed a year ago...

At today's 2024 Baidu World Conference, Baidu founder Robin Li unveiled two new AI technologies -- retrieval-enhanced Rich Graph technology (iRAG) and the code-free tool "second da". Robin Li said that the latest average daily call volume of the big model of the heart of the text has come to 1.5 billion, compared with six months ago (200 million calls) increased to 7.5 times. The illusion problem is one of the biggest limitations of the wide application of the big model, and the technology behind the solution to the problem of text generation illusion is RAG, i.e., Retrieval Augmentation. Robin Lee said that the biggest change for the big model industry in the past 24 months...

November 12 news, in today's 2024 Baidu World Congress, Baidu officially released Xiaodu AI glasses, said the product is "the world's first native AI glasses equipped with Chinese big model". According to Baidu official introduction, Xiaodu AI glasses with the first perspective shooting, ask while walking, calorie recognition, knowledge of things encyclopedia, audio-visual translation, intelligent memo and other functions. Specific configuration, the glasses: weight 45 grams, equipped with 16MP ultra-wide angle camera, support AI anti-shake algorithm; nominal standby endurance of 56 hours, support for more than 5 hours of continuous listening ...

Xiaodu announced that it will release "AI new species" at the Baidu World Congress on November 12, and this new product release will be another "new category exploration" of Xiaodu in the field of AI + hardware. At present, Xiaodu's entire product line has been equipped with large model capabilities. I learned that the official has not yet detailed what kind of new products will be released. However, earlier today, Bloomberg reported that Baidu is preparing to launch a built-in Xiaodu AI assistant smart glasses, and plans to show at the 2024 Baidu World Congress on November 12 in Shanghai. The source revealed that this...

At the 2024 Baidu Cloud Intelligence Conference, the results of the second season of the "Wenxin Cup" entrepreneurship competition were announced. Baidu founder Robin Li pointed out in his award speech that the intelligent body is his most optimistic direction of development of AI native applications, equivalent to the website in the PC era and the account in the era of self-media. Robin Li said that the most obvious feature of the intelligent body is that the threshold is low enough that anyone can get started, but at the same time the ceiling is high enough to make very complex and powerful applications. In the future, there will be millions of intelligent bodies, forming a huge intelligent body ecology. In this year's Wenxin Cup Entrepreneurship Competition, there are more than 60%...

Wenxin Intelligent Agent Platform (formerly Lingjing Matrix) is an intelligent agent building platform based on the Wenxin model launched by Baidu. It allows developers to quickly create intelligent agents through natural language interaction in a zero-code or low-code manner. The platform aims to lower the technical threshold so that ordinary users can also become intelligent agent developers, thereby promoting the widespread application and innovation of intelligent agent technology. By integrating Baidu's search, hardware, content and other resources, the Wenxin Intelligent Agent Platform provides developers with an integrated solution from development to distribution, operation and even monetization, and is committed to creating an intelligent agent ecosystem where everyone can participate and share benefits. Wen…

Baidu announced that its artificial intelligence application "Wenxin Yiyan" App has undergone a major update, launching a new version 4.0.0 and officially changing its name to "Wenxiaoyan". Wenxiaoyan App is based on Baidu's Wenxin big model, providing users with a variety of AI capabilities such as search, creation, and chat. As a special upgrade version for the first anniversary, Wenxiaoyan App has added five core scene capabilities, including asking questions, chatting, writing articles, drawing pictures, and assigning tasks. To celebrate this upgrade, Baidu announced that it will open the use of Wenxin 4.0 big model to all users for free in September. For new users or those who have not yet become...

Baidu announced today that it will jointly launch the "2024 Baidu Search Wenxin Intelligent Agent Innovation Competition" with NVIDIA, with 2 major tracks and 5 competition questions. Attached are the specific tracks: The task of the technical track contestants is to develop an LLM-based agent, given a large set of tools, to intelligently orchestrate and schedule these tools to answer complex questions in the open domain. The contestants of the intelligent agent application track need to create a life and entertainment intelligent agent based on the Wenxin intelligent agent platform around people's daily life needs, and provide users with new and innovative gameplay or leisure services. Intelligent agents should…

At the 28th Baidu hackathon on August 18, Baidu founder, chairman and CEO Robin Li suggested that young people do not have to study AI. Robin Li explained that the reason why AI has such great potential is that it may have a great impact on any industry. Just study the major you like and learn some AI-related things. When talking about campus recruitment, Robin Li said that the most important thing is the self-motivation of fresh graduates. Fresh graduates have all experienced the stage of "not knowing", but self-motivated people learn quickly, have great potential, and progress faster than others. According to...

Baidu officially announced "a special colleague" who was promoted to AI architect in just one and a half years. In fact, this colleague is Baidu's AI product - Wenxin KuaiMa, also known as Baidu Comate in English. According to the official introduction, since officially "joining" Baidu in April 2023, Wenxin KuaiMa's performance has been amazing. In August 2024, he was successfully promoted to "AI architect" and announced that he became the first AI architect in the Internet industry. Wenxin KuaiMa's work focuses on intelligent code assistants. Based on the Wenxin big model,...

At the 2024 Baidu Health Industry Ecosystem Conference, Baidu Health launched a series of large-scale model application products, aiming to use AI technology to solve the "impossible triangle" problem in the medical industry, namely "being able to see a doctor", "being able to afford to see a doctor" and "being able to treat the disease well". The following are the five large-scale model application products released by Baidu Health: Health Intelligence Family: For ordinary users, it provides AI accurate doctor search, AI medical report interpretation, AI medication assistant, AI skin detection, AI sleep assistant and other services to meet the health needs of users when they do not have to go to the hospital, and supports multi-modal interaction. Online medical Copilo…

Baidu recently launched an AI digital human social APP called "Wen Xiaoyan". This application is based on the advanced Wenxin Big Model technology, allowing users to communicate and interact with AI virtual characters in real time and establish emotional connections. Wen Xiaoyan App uses simulated digital human images to bring users a more real and natural interactive experience. After entering the Wen Xiaoyan App, users can find their favorite digital human chat partner by sliding up and down in the "Discover" function bar. Each AI digital person provides unique chat services, and he...

PaddleOCR v2.8.0, a text recognition development kit based on the PaddlePaddle deep learning open source framework, has released a milestone update. This version introduces cutting-edge OCR technology, including the champion solutions of the PaddleOCR Algorithm Model Challenge, such as the scene text recognition algorithm SVTRv2 and the table recognition algorithm SLANet-LCNetV2, setting a new standard in the OCR field. At the same time, the project structure has been deeply optimized, and non-core modules have been moved to the new warehouse, making the project more focused on OCR core technology. In addition, it has solved the problems including updating Backbo…

The official WeChat account of "Baidu Smart Cloud" announced that the Baidu Smart Cloud (Wuzhen) AI Data Industry Base, jointly built by Baidu Smart Cloud and Tongxiang Wuzhen Big Data High-tech Industrial Park, was officially launched today. The base is located in the "Direct to Wuzhen" Industrial Park in Wuzhen Town, Tongxiang City. The two parties will work together to create an AI basic data element circulation and trading gathering place, injecting new momentum into the development of Wuzhen's big data and artificial intelligence industries, with the goal of "leading the country". It was learned from Baidu Smart Cloud that Wuzhen has a sound data element industry, which will be combined with Baidu Smart Cloud Qianfan large model, autonomous driving and other technologies to achieve local...

According to Baidu's official news, Baidu Wenxin Agent Platform (AgentBuilder) has opened Wenxin Model 4.0 for free. When developers create agents on the Wenxin Agent Platform, they can flexibly choose Wenxin Model 3.5 or 4.0. Wenxin Model 4.0 was released in October 2023, achieving a comprehensive upgrade of the basic model, with improvements in understanding, generation, logic and memory capabilities. At the 2024 World Artificial Intelligence Conference held on July 4, Baidu founder, chairman and CEO Robin Li said that…

At the main forum of the 2024 World Artificial Intelligence Conference on Industry Development, Baidu founder Robin Li delivered a speech. He stressed that in the era of artificial intelligence, it is more important to develop "super-capable" applications than to pursue "super apps" with 1 billion daily active users (DAU). Robin Li believes that we should go beyond the thinking mode of the mobile era, avoid falling into the "super app trap", and realize that the definition of success should not be limited to the number of users. Robin Li proposed that the intelligent agent is the simplest form of developing AI applications and is also...

At the 2024 WAVE SUMMIT Deep Learning Developer Conference, Baidu's Chief Technology Officer Wang Haifeng officially released the Wenxin Model 4.0Turbo version. This new version has significant improvements in speed and effect, and has been launched on the official website yiyan.baidu.com. The response speed of Wenxin Model 4.0Turbo has been significantly improved. It only takes more than 20 seconds to generate a long article of more than 1,000 words, and the content quality is high. In terms of effect, the new model performs well in multiple application scenarios. It can quickly obtain information, has high requirements for the factuality and accuracy of information, and…

The Baidu Smart Cloud Qianfan AppBuilder product team released version v0.5.5, which brings a series of new features and optimizations. The most eye-catching one is the newly added long document content understanding component, which can parse documents up to 100,000 words, support information retrieval, summary and text analysis, and greatly improve reading and analysis efficiency. At the same time, the new version also introduces the digital human Agent, supports the configuration of 3D digital images, and makes the application interaction more realistic and immersive through voice dialogue and language-driven lip shape, expression and action. The application distribution channel has also been expanded, and now supports publishing applications...

Baidu AI Assistant is a multifunctional smart life assistant tool carefully developed by Baidu. It uses advanced artificial intelligence technology to provide users with a variety of Baidu AI Assistant services - Baidu AI Inspiration Center, open and use, multi-application scenario AI assistant, aiming to improve users' daily life and work efficiency. Baidu AI Assistant has multiple functions such as intelligent question and answer, search, content creation assistance, knowledge reasoning, mathematical calculation, image processing, etc. It can answer various types of questions from users, provide accurate and comprehensive information, and help users easily process pictures and perform mathematical calculations. In addition, Baidu AI…

At the Baidu WAVE SUMMIT Deep Learning Developer Conference held today, Baidu launched its intelligent code assistant - Baidu Comate. Baidu Comate is based on the Wenxin Big Model, combined with Baidu's years of programming site big data and external excellent open source data. It has multiple product advantages such as code intelligence, rich scenarios, value creation, and wide application. It can realize the scenario application form of "helping you think, help you write, and help you modify". Baidu said that among the tens of thousands of engineers in the company, 80...

Starting from June 20, Baidu Smart Cloud Xiling Digital Human Platform will usher in a major innovation and upgrade. This upgrade not only realizes the efficient and low-cost generation of exquisite 2D/3D digital humans, but also realizes matrix-style connection in multiple scenarios such as live broadcast, short video, and dialogue, which comprehensively improves the user experience. Whether it is corporate IP creation, cultural and tourism IP design, or entertainment IP development, the Xiling platform can provide one-stop intelligent, convenient, and easy-to-use digital human services. The newly upgraded Xiling platform has demonstrated remarkable capabilities in digital human generation. Users only need to enter a simple sentence describing the portrait style, and the system will…

The University of Hong Kong and Baidu jointly released the first smart city big model UrbanGPT, which has triggered a major breakthrough in the field of spatiotemporal prediction technology. The importance of spatiotemporal prediction technology is becoming increasingly prominent, focusing not only on the flow of traffic and people, but also covering multiple dimensions such as crime trends. However, due to insufficient urban data, traditional spatiotemporal prediction models are limited in terms of accurate prediction. The emergence of UrbanGPT fills this gap, and its strong generalization ability makes it show great potential in the field of zero-shot learning. Challenges faced by UrbanGPT include scarce labels and high training costs…

On May 30, 2024, Baidu Group held the Mobile Ecology Vientiane Conference in Suzhou. At this technological event, John He, Senior Vice President of Baidu Group and General Manager of Mobile Ecology Business Group, elaborated in depth on the core position of intelligent bodies in today's society, emphasizing that: "Intelligent bodies are not only the new engine of productivity, but also the key for everyone to amplify their own abilities, grasp the dividends of the times, and become a super-individual." At the conference, a number of Baidu's core businesses, including search, Wenxin Intelligent Body Platform, Wenxin Yiyin APP, e-commerce, library, and Baidu APP, demonstrated the capabilities based on the big model and intelligent body...

Recently, Xiaodu Technology, a subsidiary of Baidu, officially launched the first learning machine Z30 based on the Wenxin model, which has been fully sold on the JD platform. This learning machine is based on the core concept of "redefining AI teachers". It aims to help children actively learn through advanced AI technology, while allowing parents to enjoy a super worry-free experience of full companionship. The version with 8GB RAM + 256GB storage space is priced at 6,694 yuan. The Xiaodu Z30 learning machine uses a 14-inch 2880x1800 resolution 90Hz soft light paper-like screen, which brings users an excellent visual experience. The screen supports P3 wide color gamut, A…